calculate the return between selected time points only

Hi,

suppose I have a time series with stock prices and this variable is called "price" Now I have a second variable which is called "PSC1". This variable has n different elements which define the position of n stock prices in variable "price".

Now I want to calculate the simple returns between the selected stock prices, e.g. price(n)/price(n-1).

With my code below I always receive the famous error code:

"In an assignment  A(I) = B, the number of elements in B and I must be the same."

Thank you for any advice!

%%%%%%%%%%%%%%%%%%%%%%%%%%%

    for n = PSC1(1:end)
        TF(n)=price(n+1)/price(n)
    end

%%%%%%%%%%%%%%%%%%%%%%%%%%%%

* So PSC1 has got twenty numbers, something like 52, 54, 102... Which are single positions in the variable "price" which has a length of 800 or smth. TF is supposed to be a variable which contains 20 returns

Answer 1

[Orion]

Hi,

in your code, at the first iteration, you're trying to put a value in the 52th position of TF, which only have 20 elements => Problem !

you should do something like

for i = 1:length(PSC1)
    TF(i) = price(PSC1(i)+1)/price(PSC1(i));
end

[Jan Morawietz]

Thank you - this code works

In my original code, I did predefine TF as zerovector with same length as price. Now I understand that this pre-format is irrelevant for the iteration steps.

[Orion]

Warning / advice

You always must predefine your arrays when you know the size they will have.

- your algorithm will be faster and more readable.

- this allows you to debug errors like the one you got.